Message transmission methods of known type are generally implemented using a messaging computer program implanted in memories of computers connected to the network. By means of the messaging program, the user of the sending computer creates a message comprising both a header containing transmission information such as the address of the destination computer and the subject of the message, and also a body in which the user places the text constituting the information content of the message. The message is then sent after the sending computer has automatically filled out the transmission information, for example by adding the identity of the sending computer and time-and-date stamp data. Received messages are recorded in the destination computer so as to appear in an in-box of the messaging program. Unfortunately, when the number of received messages is large, the time required by the user in order to deal with them (i.e. to read them, to file them, to delete them, to reply to them, . . . ) is relatively long. In order to automate the filing work, the user can parameterize filters in the receiving messaging program enabling different received messages to be put into particular spaces as a function, for example, of the name of the sender of the message, of its subject, . . . . Nevertheless, when a large number of received messages all deal with the same subject or all come from the same person, the time required by the user of the receiving computer to deal with received messages remains long.
Furthermore, it can be advantageous, e.g. in the context of a work group, for the various users of computers connected to a network to use the same method of filing messages on the basis of the same classification. This is possible only providing the content of the computer of each of the users is verified periodically and the received messages are reorganized where necessary, which is not conceivable under most circumstances, in particular given the large number of computers connected to the network (several hundreds to several thousands in large businesses), the geographical distances between them, and the reticence of users.
Furthermore, when a group of users of a network, such as the employees of a business or individuals connected to the Internet, seek to share information interactively by structured interchanges, they need a dedicated server and they must form a so-called “community” thereon. To do this, the user originating the community must establish a list of members of the community and rules defining how members access the server and the information it contains. Setting up such a community also requires its members to install software on their own computers in order to interpret the protocol for communication with the server, to read and write information for sending thereto, for downloading information recorded thereon, . . . . The server to which each member can make a connection using the above-mentioned software serves to administer the community by managing member access, and it centralizes the information sent by members, filing it and storing it hierarchically. The server is also required to control the facilities allowed to members for acting on the information to which they have access. Setting up such a server is expensive, particularly because of the cost of the server if it needs to be purchased, the cost of the software using the protocol specific to the server used, costs involved with managing the server, costs accrued by members of the community connecting to the server, . . . . It is also time-consuming to set up a server and it is not very interactive in use since any modification relating to the members of the community, to their modes of access, or to the facilities they have for acting on the information requires action to be taken at the server, and this must be done by qualified personnel. Thus, the use of such servers is particularly rigid, and that goes against users' desire for flexibility in interactively sharing the information contained in the server. Furthermore, when the volume of information on the server becomes too large, the oldest information is removed. This removal is performed at the discretion of the management of the server without necessarily referring to the members of the community. This reduces the overall memory of the community.
Such interchanges of messages are regularly undertaken by work groups or by discussion forums or “news groups”.